Hard to answer without a picture and any details. My only guess is that the bound RNA is protected from degradation by RNAse.

Hi,

picture of gel would be very helpful.

According to your description there might be few possible problems:

1) An obvious one: your highest protein concentration does not exceed RNA concentration.

2) RNA-protein complex is not very stable. You can run it at low voltage (100V in miniProtean3 device for expl) and at low temperature.

3) Some ions are needed for binding. If so add them not only to the binding buffer, but also to the running buffer. Also consider removing EDTA from running buffer.

4) Your RNA sample is a mix of the one which is supposed to be bound, and the one which is not. Therefore even low amounts of protein would bind all of the good ones, and the rest would stay free.

5) The wells were not washed with buffer. This would explain the randomness: "sometimes it appears, sometimes not"

6) Incubation of RNA-protein mix was not long enough. Usually its at least 30' on ice (unless its a thermostable protein).

7) The gel percentage is not selected properly. If the complex is too big and the gel is too dense - it will not enter fully or at all.

8) pI of your protein is very high (>9) and its much bigger than RNA. This can result in the protein refusing to to migrate to the anode.

There could be more - EMSA is tricky and has to be optimized for each complex.

Hope this helps. Good luck :)